Current techniques for the screening and risk assessment of prostate cancer, as a prerequisite to surgical biopsy, are based upon measurements of either individual serum biomarkers, or of expression of individual genes in circulating malignant cells. These techniques possess a number of limitations, including lack of specificity and accuracy in diagnosis, and a lack of detailed prognostic information. The resultant high numbers of false positive diagnoses lead to numerous unnecessary surgical biopsies. There is evidence that individuals with prostate cancer and other forms of malignant disease exhibit immune responses that can be detected at the level of altered gene expression in leukocytes circulating in peripheral blood. Quantitation of the mRNA transcripts in circulating leukocytes of a number of individual genes has demonstrated an association between gene expression level and the presence of a tumor. Furthermore, it has been shown that serum levels of individual proteins, which exhibit a degree of correlation with differential gene expression in leukocytes, provide some information on tumor stage. We have initiated a pilot study to measure gene expression in leukocytes and have produced initial data that concurs with previous reports. We have further documented the use of algorithms to classify prostate cancer patients and healthy controls. Under Specific Aim One, we will collect and process blood from prostate cancer patients and healthy control subjects. We will then employ microarray technology to measure simultaneously the expression levels of up to 14,000 genes transcribed in leukocytes derived from the blood of prostate cancer patients and control subjects, and employ data analysis algorithms to determine patterns of gene expression specific for each subject group. With this technology we propose to investigate our central hypothesis: that individuals suffering from prostate cancer exhibit a conserved pattern of gene expression levels in their peripheral blood leukocytes, which is distinct from the pattern of expression in peripheral blood leukocytes from control subjects. Under Specific Aim Two, we will test the further hypothesis that cancer patients with prostate tumors at different histological grades will yield distinct expression signatures that reflect the biological stage and aggressiveness of the tumor, and that can thus be employed to differentiate among tumors at different pathological stages. We believe that the diagnostic technique we propose to develop may ultimately form the basis of a clinical assay that will, with a minimum of patient discomfort, have the capacity to identify men with prostate cancer, and also provide important stage-specific information.